Dynamic properties of simulated brain network models and empirical resting-state data

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• Brain network models (BNMs) simulate whole-brain activity, like resting-state fMRI (rs-fMRI).
• Comparing simulated complex brain activity to empirical data remains challenging.
• Previous methods often relied on simple metrics like functional connectivity.

Purpose of the Study:

• To adapt and apply dynamic analysis tools, typically used for empirical rs-fMRI, to simulated BNM data.
• To investigate how different dynamic analysis metrics differentiate between various BNMs.
• To identify which analytical approaches best capture the complexities of simulated versus empirical brain dynamics.

Main Methods:

• Applied diverse dynamic analysis tools to simulated brain activity from BNMs.
• Compared results from dynamic analyses to structural connectivity inputs and model mathematical descriptions.
• Evaluated the contrast between BNM simulations and empirical rs-fMRI data using these tools.

Main Results:

• Some BNM properties correlated with their shared structural connectivity.
• Other dynamic properties reflected the mathematical underpinnings of the BNMs.
• Dynamic analyses focusing on temporal signal patterns showed greater contrast between BNMs than spatial coordination metrics.

Conclusions:

• Dynamic analysis of temporal signal patterns offers a more sensitive approach to differentiate between brain network models.
• This method provides a better way to compare simulated brain activity with empirical data.
• Findings will aid in refining and developing more realistic whole-brain activity simulations.